Should students vote individually, should you have a whole class discussion? (#clicker series)

by Stephanie Chasteen on April 12, 2013

hand-on-clicker-100pxI’m continuing my series of posts on the literature related to clickers and the use of Peer Instruction.  Catching up on some good articles!

This post regards an article from 2003, “Peer Instruction vs Class-wide Discussion in Large Classes:  A comparison of two interaction methods in the wired classroom,” D. J. Nichol and J. T. Boyle, Studies in Higher Education, 28(4), 2003.

This study addresses the fact that instructors are using clickers in a wide variety of ways.  This has been documented before by Henderson and Dancy (PDF) in the physics community, when instructors who indicated that they used peer instruction often dropped what are considered critical components of the method, such as the use of conceptual questions or peer discussion.

This project  asked mechanical engineering students about their opinions and experiences regarding two of the core aspects of peer instruction — the peer-peer discussion portion of clicker questions, and the whole-class wrap-up after the votes are in.  The lecture hall is modified to accommodate interactive teaching methods (i.e., students sit in groups of four), and clicker groups remained the same during the course.  Students were introduced to peer instruction early in the semester, with some good attention paid to motivating them to participate in the process.  Questions were quite conceptual, and based on common student difficulties.  Students voted on their own, discussed in groups, and then re-voted.  The student vote on these questions was used as a trigger for peer group discussions as well as teacher-facilitated whole class discussions, resulting in what seems to be quite a flexible teaching arrangement.

The Comparison:  Mazur vs PERG sequences

The researchers attempted to compare student perceptions of two possible sequences, one dubbed the “Mazur” method, popularized by Eric Mazur in his book Peer Instruction.  The other, the PERG sequence, is that used by the University of Maryland group.    I’ve highlighted the places where the two sequences differ

Mazur sequence (Used in weeks 9 and 10)

  1. Pose question
  2. Individual reflection and individual vote
  3. Show individual vote histogram
  4. Peer discussion
  5. Students re-vote
  6. Show histogram of re-vote
  7. Instructor summarizes and explains

PERG (Used in weeks 6 and 7)

  1. Pose question
  2. Peer discussion
  3. Students vote
  4. Show histogram of vote
  5. Class-wide discussion facilitated by instructor
  6. Instructor summarizes and explains

So, the Mazur method includes an individual vote, but not a whole-class discussion, at least typically.  (Note that here at the University of Colorado Boulder I would say we typically follow the PERG method).

Survey and Interviews

The researchers used a survey as well as student focus groups (25% of the class) to examine student perceptions.  Focus groups were separated by gender, and the class lecturer did not take part in the interviews or analysis.  The survey was developed based on the outcomes of the focus group interviews, and were intended to determine how widespread some of the ideas from interviews were across the class population.  Lastly, they used a “critical incident questionnaire,” where students were asked to reflect on a recent interactive session, and indicate where they felt most engaged.  I really like the mixed-methods used in this study!


As in previous studies, students said that they learned from these teaching methods.  (The authors report on these results in more detail elsewhere).

With respect to the peer discussion portion of the methods, students indicated that hearing the arguments put forward by other students helped them to change their conceptions of the material.  Many comments indicated that they also felt more comfortable telling a peer that they didn’t understand something, and that it was easier to understand an explanation from a peer than from an instructor, since the peer was closer to their level of understanding and used more accessible language.

On the “critical incident questionnaire,” students reported that they felt most engaged when they were interacting and discussing problems with other students.  So, there is ample evidence that student/student discussion is a powerful force for learning and engagement.

Here are some of the most interesting findings from the survey:

  • It is better to answer the concept question individually before discussing the same question in small groups:  82% agree
  • Group discussion after making an individual response leads to deeper thinking about the topic:  90% agree
  • It is better to start with a small group discussion before making an individual response:  13% agree
  • The best interaction method depends on the difficulty of the question:  43% agree
  • A class discussion using a microphone is an important aspect of the class:  40% agree
  • I like having to explain answers to the questions using the microphone:  10% agree
  • It is important that the teacher clearly explains which is the right answer and why after a class discussion using the microphone:  100% agree

So, some of the important take-home messages are:

1.  Students prefer to think about a question individually first.

Students felt they’d be more likely to be passive, and influenced by confident peers, if they haven’t had time to think about it on their own first.  They also felt they got more out of peer discussion if they had already determined what they thought about the question.

This is interesting, and matches with what I’ve often thought might be the case.  However, at Colorado, we often drop this individual vote in the interest of time.  Perhaps it’s OK to not sacrifice the good to the perfect, however?

2.  Students were mixed about class-wide discussions

Students preferred the peer discussion to the whole-class discussion.  While students liked hearing explanations from students outside their group, they were mixed on how important they thought this was (only 40% agreed).  Some also indicated that they were more attentive if they knew they might be called upon to share their response.  However, they felt there were many drawbacks to the class wide discussion.  It took too much time, it was easy to get off topic, and sometimes the discussion and unclear arguments caused further confusion.  Students also found it anxiety-provoking to be called to the microphone.   At CU, our classes are small enough that no microphone is required, and we ask for student volunteers — so I wonder how these results might be different in that format.

However, students did indicate that it was most helpful to use this whole class discussion when the histogram showed that the class was clearly divided between two or more answers.  Indeed, this is when we advocate using whole class discussion — when it has the most potential to clarify the reasoning behind different popular answers.

Note that students were unanimous that it is important for the instructor to clarify the correct answer and reasoning — a finding that mirrors that found in Smith and Knight on the importance of combining peer discussion with instructor explanation.

The authors have an interesting suggestion why the whole-class discussion might be of less value:  In these discussions, students are not generating their own mental models, and so there is less conceptual conflict to resolve during this discussion.

Take-home messages

So, based on the results of this study, it is best to:

  1. Have students vote individually prior to peer discussion (unless the question is likely to be too difficult for students to think productively about it on their own
  2. Use peer discussion
  3. Use whole-class discussion only when there is more than one popular answer
  4. Keep the whole-class discussion focused and within a reasonable time limit, and
  5. Clarify the correct answer and reasoning at the end

The authors remark on a fact that continues to be heartening to me — in these sorts of studies, over and over, students and instructors both provide ideas about learning which dovetail remarkably well with the research literature.


Leave a Comment

Previous post:

Next post: